The global population is ageing at a fast pace. In 2050, the proportion of the world’s population over 60 years is expected to reach 22%, almost doubling the 12% of 2015 (1). The World Health Organization (WHO) defines elderly the subjects aged beyond 65 years (65y+), a highly heterogeneous group of people in terms of performance status, comorbidities and vulnerabilities.

The variability of elderly patients poses significant challenges to the clinical practice of oncology (2, 3), including the context of rare cancers, where few clinical trials are available and study populations are less numerous (4). Clinicians have few tools to estimate the risk/benefit balance of treatments for elderly patients, since an age subgroup-specific analysis is unlikely feasible. However, the differences in survival outcomes from rare cancers in the United States and Europe are more evident for age group rather than for cancer type, with the 65y+ subjects at high risk of poor outcome (5).

The level of clinical challenge is major for patients with rare tumors requiring a multidisciplinary approach, as in the case of head and neck cancers (HNCs). In patients with HNCs, both age and comorbidities influence the overall survival (OS), possibly because of advanced tumor stage, inability to perform multimodal treatments and/or non cancer-related causes of death (6). Moreover, the consideration of chronological age and mild/moderate comorbidities may result in the selection of substandard treatments, which are associated to lower OS and cancer-specific survival, in comparison with the state-of-the-art (7). However, HNCs encompass a heterogeneous group of cancers with different histologies and clinical courses. This review is focused on salivary gland cancers (SGCs) in elderly patients, including the most recent epidemiological findings, the current treatment approaches and the future perspectives.

SGCs are rare cancers, accounting for less than 5% of all malignancies of the cervicofacial region. The WHO Global Cancer Observatory (Globocan) reported 53.583 new cases of SGCs diagnosed in 2020 worldwide, 43% occurring in the elderly, and causing 12.339 cancer-specific deaths, with a male-to-female ratio of 1.3:1. In the next two decades, the new diagnoses in the elderly age group are expected to account for 80% of the total SGCs diagnoses; similarly, SGCs in this group are expected to cause the 88% of all SGC-specific deaths (8). The impact of SGCs in the elderly is going to become a crucial health issue, thus research efforts should be encouraged to identify the main risk factors and the most effective therapeutic strategies for this multifaceted population.

The ICARE study, a multicenter, population-based, case-control study recently conducted in France on 73 SCGs and 3555 controls, reported that the main risk factors for SGCs were related to a previous history of cervicofacial radiation therapy (RT), either for HNCs (odds ratio, OR = 31.74, 95% CI 2.5 – 405.2) or hematological cancers (OR = 5.1, 95% CI 0.6 – 46.2), and professional exposure to chemicals, such as metals in the plumbing industry, electrical equipment and nickel compounds/alloy. In the ICARE study, the mean age at diagnosis was 56.9 years, with 27% of the population aged more than 62 years (9). Another case-control study conducted in Japan showed that heavy smokers were at higher risk of developing SGCs, compared with never smokers (OR = 3.45, 95% CI 2.06-12.87; p < 0.001); 43.7% of the study population was more than 60 years of age (10).

On the basis of the current literature, it is unknown whether the onset of SGCs in the elderly population could be sustained by risk factors that are different from those of younger patients. However, some differences can be observed not only in the survival outcomes, but also in the histotype-specific incidence. The clinical management of elderly patients has improved in the last decades, due to new clinical tools of geriatric assessment. The Comprehensive Geriatric Assessment (CGA) is a method for identifying elderly patients at risk of life-threatening events during oncological therapy by analyzing several domains (functionality, nutrition, cognition, psychological state, social support, comorbidities, medication review, and geriatric syndromes). It can predict functional decline and also be used to adapt cancer treatment, as demonstrated by the ELCAPA study (11). Since CGA is a time-consuming tool, in the last decade various screening instruments were adopted, in order to refine the selection of vulnerable patients who could benefit from a CGA. The G8 questionnaire, based on seven items from the Mini Nutritional Assessment (MNA) and age as the 8^th item, uses a scoring system from 0 to 17, where a result below the cut-off value of 14 identifies those patients who should be addressed to CGA. G8 has been validated in multicenter cohort studies (12, 13) and also in a population of elderly HNSCC patients treated with chemo(radio)therapy (14). G8 has high sensitivity, but a proportion of patients with G8 scores ≤14 has no major vulnerabilities detected by the CGA. Moreover, the level of G8 specificity may vary according to the primary tumor sites, as demonstrated by the ELCAPA-02 study, which included few patients with HNC (n=4) (15). An optimized version of the G8 was recently proposed, including six independent predictors for abnormal CGA: weight loss, cognition/mood, performance status, self-rated health status, polypharmacy (≥ 6 medications per day), and history of heart failure/coronary heart disease (16).

Focusing on the treatment of patients with SGCs, the recently released ASCO Guidelines recommend taking clinical decisions in the context of a multidisciplinary tumor board, focusing on histology, disease burden and site of tumor deposits, potential treatment-related toxicities, patient’s overall health, comorbidities and function (17). Comorbidities are the Achilles’ heel of elderly patients with cancer, however, patients with SGCs have less comorbidities compared to those with other HNCs. In a retrospectively analyzed cohort of 666 patients with SGCs, OS – but not disease-free survival (DFS) – was influenced by the Adult Comorbidity Evaluation index (ACE-27). The ACE-27 scoring was affected by age and gender, possibly influenced by lifestyle. In that study, patients with comorbidities were more likely to receive a non-surgical treatment, i.e. exclusive RT, or no treatment at all, achieving a worse outcome. The impact of comorbidities was not related to the histological subtypes, except for the squamous cell histology. Interestingly, DFS was not influenced by comorbidities, indicating that a proper treatment should be delivered as much as possible, even in elderly patients with comorbidities (18).

Another study focused on the influence of age and comorbidities on the outcome using the Age-Adjusted Charlson Comorbidity Index (ACCI) scoring system on a series of 109 patients with a median age of 69 years, treated for major SGCs. Comorbidities, but not age, were an independent prognostic factor for both OS and disease-specific survival (19). A Danish study performed in 871 patients with SGC treated between 1990 and 2005, reported a poorer survival in the population group over 70 years old (n=282), possibly explained by the more advanced disease stages, poorer performance status at the time of diagnosis, more high-grade histological subtypes such as adenocarcinoma not otherwise specified (NOS) and carcinoma ex-pleomorphic adenoma (ca ex-PA). Despite these aggressive clinical features, elderly patients received surgery plus RT only in 45% of cases, while 54% received a suboptimal active treatment, either surgery alone (38%) or exclusive RT (10%), or best supportive care (6%) (20).

Malignant tumors of the major salivary glands (MSG) can be found in 15-32% of parotid, 41-45% of submandibular and 70-90% of sublingual masses. The minor salivary glands (mSG) are located beneath the mucosa of the oral cavity, palate, paranasal sinuses, pharynx, larynx, trachea and bronchi, mostly concentrated in the buccal, labial, palatal, and lingual regions. Almost 50% of the tumors arising from mSG are malignant (21). The staging of MSG is currently based on the 8^th edition of the American Joint Committee on Cancer/Union for International Cancer Control (AJCC/UICC) tumor/node/metastasis (TNM) System, while mSGCs are staged according to the AJCC/UICC system for the primary site (22). In patients with a suspicion of SGC, the ASCO Guidelines recommend performing an imaging workup including neck ultrasound, computed tomography (CT) with intravenous contrast, and/or magnetic resonance imaging (MRI) of the neck and primary site (Recommendation 1.1) (17). An MRI with diffusion sequence of the neck and skull base is recommended in case of suspicious perineural spread and/or skull base involvement (R1.3), and an MRI brain scan could be considered in case of high-grade SGCs or if suspected meningeal spread. A CT scan should be performed in case of possible involvement of the adjacent bone (R1.2). An FDG-PET/CT scan from the skull base to mid-thighs may be performed for tumors with high-grade features, while in low-grade histological subtypes it could deliver false negative results (R1.4). For locoregionally advanced cases, a contrast-enhanced chest-abdomen CT scan is recommended to complete the clinical staging. Intracranial metastases are infrequent at diagnosis; however, they may occur especially in case of high-grade SGC (23).

Beside the multiplicity of anatomical subsites, SGCs harbor a wide heterogeneity of histologies, which are associated to different age incidence, clinical behavior, treatments and prognosis. Each glandular segment can be the site of origin of SGCs ( Figure 1 ). The WHO 2017 Classification of Salivary Gland Tumors (4^th Edition) classified more than 20 malignant subtypes of epithelial SGCs, including a new entity, the secretory carcinoma (SC), formerly known as mammary analogue SC (MASC) (24, 25). Moreover, the recently released Armed Forces Institute of Pathology (AFIP) Atlas of SGC Tumor Pathology provides an indepth analysis of the morphological and molecular features of both benign and malignant salivary gland tumors (26).

Recently, the molecular landscape of SGCs has been explored by comprehensive genomic profiling, opening this complex disease to the possibilities of targeted therapy (27, 28). Molecular alterations found in SGCs include amplifications, mutations or rearrangements of transmembrane receptors (ERBB2, FGFR, PDGFR, RET), mTOR pathway (PIK3CA, PTEN) and MAPK pathway (BRAF, HRAS), DNA repair (BRCA1/2), cycle cell regulation (CDKN2A/B, SMARCB1) and activation of androgen-responsive genes by the androgen receptor (AR). Interestingly, the upregulation of ERBB2/PIK3CA pathways is more frequently observed in high-grade than in low-grade SGCs (27). However, certain histotypes can encompass both low-grade and high-grade forms, as in the case of polymorphous adenocarcinoma (PAC), previously known as polymorphous low-grade adenocarcinoma (PLGA); in the last 2017 WHO classification of Head and Neck tumors, both the classic variant of PLGA and aggressive cribriform adenocarcinoma of minor salivary glands (CAMSG), were incorporated under the PAC category. Interestingly, both the variants harbor in the majority of cases an alteration of PRKD genes codifying for Serine/Threonine-Protein Kinase D1, most frequently mutations in the classic PAC variant and rearrangements in CAMSG (29). Certain key rearrangements, such as MYB-NFIB, NR4A3, PLAG1, ETV6-NTRK/RET, CRTC1-MAML2 are typical of certain histotypes, and may help pathologists in the diagnosis of challenging cases (30–36) ( Table 1 ).

The most frequent high-grade histotypes are mucoepidermoid carcinoma (MEC), salivary duct carcinoma (SDC), adenocarcinoma NOS (AD-NOS) and carcinoma ex-pleomorphic adenoma (ca ex-PA). Low-grade histotypes include adenoid cystic carcinoma (AdCC), acinic cell carcinoma (AcCC), myoepithelial carcinoma (MyoEpi), and SC (previously MASC). However, both AdCC and AcCC may develop a high-grade transformation by dedifferentiation, and this phenomenon has been described also in elderly patients (37, 38) ( Table 1 ).

Tumor grading is directly related to the disease aggressiveness, and it is of utmost interest to capture the age distribution of the commonest types of SGCs. As reported in the AFIP Atlas, certain histotypes are more frequently diagnosed in the 6^th decade and beyond, especially SDC, AcCC with high-grade transformation, MyoEpi, basal cell adenocarcinoma, primary squamous cell carcinoma (SCC), large cell undifferentiated carcinoma, epithelial-myoepithelial carcinoma, and carcinosarcoma (26).

For the purpose of this review, we report the age-stratified incidence of epithelial SGCs using the Surveillance, Epidemiology, and End Results (SEER) Program database provided by the U.S.A. National Cancer Institute’s Division of Cancer Control and Population Sciences (39). We selected the cases diagnosed between 2011 and 2018 and retrieved from the updated dataset of the SEER Program based on 18 registries from 13 States (88.816.582 inhabitants). SGCs cases were defined using a combination of the International Classification of Diseases for Oncology (ICD-O) morphology and topography codes (40), as proposed by the RARECAREnet project (5). The SGCs included in the SEER database were divided into two groups:

1 - malignant epithelial tumors of the major salivary glands (MSG);

2 - malignant epithelial salivary glands tumors other than MSG (SGT).

In both groups we report the incidence and 5-year survival (%) by morphology codes, stage at diagnosis and age groups (0-65y and 65y+) ( Tables 2 – 4 ). The methods used for the calculation of the frequency (incidence) and outcome (relative survival) were provided by the SEERStat program. The number of new diagnoses made in one year for each specific population of cases (incidence) was expressed in annual rate, defined as the number of cases on 100,000 inhabitants per year. The relative survival is the ratio between life expectancy of the cohort of cases affected by cancer and the life expectancy of the population of cases. It is the closest estimation of cause-specific survival in clinical studies.

In comparison with 0-65y subgroup, in the 65y+ the incidence rates were 4 and 7 times higher for SGT and MSG, respectively. Focusing on the histotypes, AcCC, MEC and AdCC of SGT were more frequently diagnosed in the 0-64y cohort, while SCC were more common in elderly patients with MSG. Moreover, the neoplasms NOS of MSG and AD-NOS of SGT were more commonly diagnosed in the elderly, compared with younger patients (70% vs 30%) ( Table 2 ).

In the SEER dataset population, the overall 5-years survival rates were significantly better in young than in elderly subjects: the survival rates after MSG and SGT diagnoses were 82% and 89% in the 0-65y group, 61% and 81% in the 65y+ group. Interestingly, the worst outcomes were reported for elderly patients with MSG. The 5-years survival rate was lower for elderly patients in the majority of histotypes, especially in neoplasms NOS (34%), AD-NOS (57%), MEC (78%) of the MSGs and AcCC (81%) of the SGT ( Table 3 ). The differences in survival outcomes observed between the two age groups may be partially explained by a higher proportion of histotypes with favorable prognosis in the 0-65y group, compared to the 65y+ ( Table 2 ).

Almost half of SGC cases from the SEER dataset presented at diagnosis with localised disease, without significant differences between MSG and SGT ( Table 4 ). Elderly patients were diagnosed more frequently at a metastatic stage, and the 5-year survival for each stage was lower in the elderly subgroup. Focusing on the cases with available staging, the greatest difference between age subgroups could be observed for locoregional disease in MSG, where the multidisciplinary team is crucial to define an optimal management.

The role of RT in the management of SGCs mainly consists of postoperative radiation therapy (PORT) initiated after the assessment of pathologic risk factors and within 8 weeks post-surgery (Recommendation 3.7) (17), but for patients with medical comorbidities or unresectable tumors, primary RT with curative intent or – more frequently – with palliative intent are also options (Recommendation 3.10) (17). PORT is recommended for tumors with at least one of the following features, according to the ASCO 2021 (Recommendation 3.2) and the National Comprehensive Cancer Network (NCCN) guidelines v.1.2022 (75):

Target volume selection and delineation depend on the location of the primary tumors, the histologic subtype and the neck status. For parotid and submandibular gland tumors, PORT for primary tumor should include the tumor bed with a margin to cover all the surrounding normal tissue in which microscopic tumor infiltration could have spread, such as the para-pharyngeal space for deep parotid tumors, or part of the masseter muscle in case of accessory parotid infiltration. There are no guidelines for mSGCs, and the delineation should be done according to the location of the primary tumor. Several authors published guidelines on target volume delineation for parotid and sub-mandibular glands, and readers are referred to these publications for more detailed information (76–79).

In case of AdCC of the parotid gland, owing to the neurotropism of this histology, delineation of the VII cranial nerve should be done, including the skull base for nerve infiltration close to the mastoid; as AdCC cells may spread along the auriculotemporal nerve sheath, the 3^rd branch of trigeminal nerve (mandibular nerve, V3) could also be at risk of infiltration, especially for parotid tumors invading the masticator space. For submandibular gland tumors, the branches of the XII nerve and the lingual nerve (a branch of V3) should be delineated. All patients should be treated with IMRT or VMAT, and a dose of 60 Gy should be delivered in 30 daily fractions of 2 Gy (80, 81). There are no data to justify a higher radiation dose in case of positive resection margins (R1), and the benefit of concomitant chemo-radiotherapy is still unknown. The only randomized controlled trial designed to answer this issue (RTOG 1008) has closed to participant accrual and currently is in the phase of data analysis (NCT01220583). Retrospective data do not support RT implementation with chemotherapy on a routine basis (82), and the most recent ASCO Guidelines do not recommend this strategy (17).

After elective neck dissection, PORT is not recommended for pN0 cases and for patients with a single positive node without extra-nodal spread (ENE-). In cN0 patients who did not undergo neck dissection, a watchful policy is recommended for low-grade and low-stage tumors. Conversely, a prophylactic irradiation of level Ib to IV is advised for high grade tumors, SDC, SCC, adenocarcinoma NOS, and undifferentiated histologies. For patients with multiple positive lymph nodes (N+) and/or for those presenting with extra-nodal extension (ENE+), irrespective of the number of positive nodes, PORT of levels Ib to V is recommended (70, 83).

Guidelines on target volume delineation in the neck have been published (84, 85). Typically, elective dose is delivered in the range of 50 Gy in 2 Gy equivalent, whereas a dose of 60 Gy in 30 daily fractions is recommended for pN+ neck (70). Current data do not justify a higher radiation dose in ENE+ cases, and the benefit of concomitant chemotherapy is unknown. As for the primary tumor bed, intensity-modulated RT (IMRT) or volumetric modulated arc therapy (VMAT) should be used as standard irradiation techniques (86–88). When two different dose levels are used, it is recommended to use a Simultaneous Integrated Boost (SIB) approach.

In patients with unresectable or inoperable tumors due to comorbidities, exclusive RT – either with photons or heavy particles – should be considered as a treatment option. For patients who may not be able to receive curative treatment, palliative RT can improve the quality of life by controlling major loco-regional symptoms such as pain, dysphagia, dyspnea, and bleeding. There is no specific protocol for SGCs, and different RT regimens may be administered, such as 30 Gy in 10 fractions, 20-25 Gy in 5 fractions, or 40-50 Gy in 16 fractions (89). Studies focused on target volume selection, delineation, and dose level as a function of patient age/comorbidities are currently lacking. A study compared a small cohort of 29 elderly patients with major SGCs treated with chemo-radiotherapy or PORT with a matched-pair group of younger patients, without finding any difference in acute or late toxicities (90). In RT for mucosal SCC, age was not reported as a prognostic factor for the development of acute or late toxicity, therefore it is recommended to follow similar guidelines for SGCs RT in elderly as in younger patients (91), also considering that irradiated volumes are typically smaller for SGCs in comparison to other HNCs.

SGCs of the head and neck are frequently found in proximity to critical structures, such as the paranasal sinuses and the base of skull, and their local control following RT is dose-dependent. Especially in SGCs, the progress of RT techniques throughout the past decades has been achieved by increasing accuracy, and the dose escalation facilitated by high-precision technologies. As opposed to photon RT, proton and heavy ion (carbon) beams exhibit a finite penetration depth within tissue and deposit most of the energy at the end of their path (Bragg Peak), with only minimal dose deposition beyond the Bragg peak. This allows the generation of extremely steep dose gradients resulting in improved sparing of the normal tissue surrounding the target volume. In addition, carbon ions beams generate more complex DNA damage, leading to increased biological effectiveness as compared to either photon or proton RT, and making this technique ideal for SGCs treatment.

High linear energy transfer (LET) RT for SGCs has been explored early on (92, 93). Despite long-term toxicity was significant, the treatment with neutrons achieved a higher local control of disease compared with photons RT, and these early studies led to the investigation of charged particle therapy for SGCs.

Adjuvant systemic treatments are not recommended by the ASCO Guidelines outside of clinical trials, despite the fact that two retrospective studies described a potential benefit of androgen deprivation therapy (ADT) and HER2-blockade in SGCs harboring high-risk pathological features of relapse and overexpression of androgen receptor (AR+) and HER2 receptor, respectively (132, 133). Both studies reported a median age at diagnosis of 60 years (range 29–84 years in the study with ADT; 18–87 years in the study with HER2-blockade) and a prevalence of high-risk histologies, such as SDC. In the former study, 22 patients with stage IVA AR+ SDC received bicalutamide alone or bicalutamide plus LHRHa as adjuvant treatment. At a median follow-up of 20 months, the 3-years disease free survival (DFS) of ADT-treated patients and control group was 48% and 27.7% respectively, with a hazard ratio (HR) of 0.14 (95% CI 0.025 – 0.75, p=0.02). In the latter study on adjuvant HER2-blockade, 34 patients were HER2 positive (1-3+) by immunostaining (41% 3+, 38.2% 2+ and 20.6% 1+). A statistically significant survival benefit of adjuvant trastuzumab versus standard follow-up was observed only in the subgroup with high expression of HER2/neu at immunohistochemistry (IHC 3+ score), showing a DFS of 117 vs 9 months (p = 0.02) and an OS of 74 vs 43 months (p = 0.02). A trial with the antibody-drug conjugate (ADC) ado-trastuzumab emtansine (T-DM1) in the adjuvant setting of HER2-positive SGCs is ongoing (NCT04620187).

Currently, the setting of systemic treatments for SGCs is recurrent/metastatic (R/M) disease without either surgical or (re-)irradiation options. Common sites of distant metastases in SGC are the lung (49–91%), bones (40%), liver (19%), soft tissue (9%), distant nodal basins (8%), brain (7%), kidney (2%), orbit (2%), pancreas (2%) (134).

As specified in the last 2021 ASCO Guidelines, therapy initiation should be considered in the following situations (Recommendation 6.3) (17):

Notably, patients with oligometastatic AdCC and low-grade SGCs with a limited metastatic burden (i.e. ≤ 5 lesions) can be considered amenable to palliative local treatments, either by surgery (e.g. lung metastasectomy) or stereotactic body radiation therapy (SBRT) (Recommendation 6.2) (17). SBRT would be the most suitable option to fragile patients facing a high surgical risk for advanced age and/or comorbidities.

Two single-center studies focused on the setting of advanced AdCC showed that increasing age was related to worse OS: in one study of 105 patients with median age 57.3y (range 19 – 87), the hazard ratio per decade for OS was 1.2 (CI 1.1 – 1.6) (135). In another study of 88 patients with a median age of 58y (26 – 79), the group aged > 60y had worse OS compared with the < 40y group (p=.028) (136). Recently, a prognostic nomogram was proposed as a tool which may help clinicians in recognizing patients with metastatic AdCC who could benefit from a watchful waiting strategy vs active treatment (137). The study included 298 patients with a median age of 51y (range 42 – 60) in the testing cohort (n=259) and 59y (range 48 – 68) in the validation cohort (n=39). The nomogram is based on five independent prognostic factors (gender, disease-free interval and presence of lung, liver or bone metastases) that can predict the overall survival at 3, 5 and 7 years. Therefore, age was not confirmed as an independent prognostic factor for OS in this group of R/M AdCC patients.

Elderly patients with SGC carry a poor prognosis compared to younger patients. We have reported the incidence and the 5-years survival outcomes of the most common histotypes in the U.S. population, in order to assess whether a different distribution of histotypes in the two main age subgroups (< 65y and 65y+) could explain part of this survival disparity. Good prognosis histotypes and early stage at diagnosis were more common in younger patients. Conversely, advanced stage at diagnosis and unspecified (NOS) histotypes presented more frequently in the elderly. These characteristics are possibly related to the aggressiveness of disease, poor attention to early symptoms, and difficulty in accessing referral hospitals to receive an accurate diagnosis and appropriate treatment. Such features, typically found in elderly, have been recognized as independent poor prognostic factors in SGC patients, regardless of the type of treatment (17, 18). Also, the high proportion of “carcinoma, NOS” in the elderly group could be an indirect sign of suboptimal pathological diagnosis in this population.

Since a correct pathological diagnosis is the prerequisite for an appropriate treatment planning, proper pathological and molecular analyses should be carried out, both for diagnostic and therapeutic purposes, in order to avoid the generic diagnosis of “carcinoma, NOS”. As assessed in the ASCO 2021 Guidelines (Recommendation 1.6), the use of ultrasound-guided core needle biopsy (CNB) has an estimated sensitivity of 94% and specificity of 98% (204). Moreover, CNB has a lower inadequacy rate (1.2%) than FNAB (8%) and yields adequate material for ancillary molecular testing (205, 206).

In Europe, the centralization for HNCs cases is feasible and active (207). However, existing differences among countries have been reported (208). In consideration of the multiple and challenging clinical variables (e.g. comorbidities, high-grade tumors, advanced stage), an optimal care of elderly patients with a diagnosis of SGC should start with a multidisciplinary tumor board evaluation, together with a G8 questionnaire, and a comprehensive geriatric assessment should be offered whenever needed ( Figure 2 ).

Surgery is the hinge of curative pathway in SGC patients, and not less so in the elderly population. An advanced chronological age is not an absolute contraindication for surgery. The tumor resectability, the extent of surgery and the expected outcomes should be discussed. In addition, screening tools such as the Age Adjusted Charlson Comorbidity Index and G8 questionnaire have to be assessed also in this setting, in order to identify frail patients and their special needs.

Postoperative RT is generally delivered in advanced stage, high-grade or in case of high-risk pathological features. The choice of radiation type – by photons or heavy particles – should be based on clinicopathological characteristics, resection margins, and facilities availability. Toxicities should be closely monitored and early managed, with a careful and continuous supportive care during the whole treatment period. For those patients not amenable to surgery, either due to the extent of disease or comorbidities, exclusive RT with radical or palliative aim would be most reasonable choice. Based on the available data, there is no evidence to support that elderly patients undergoing particle therapy should be treated differently than younger subjects. In fact, toxicity profiles and control rates favor particle therapy over photon RT. There is a general lack of information on structured quality of life assessments in the particle therapy literature throughout all age groups, and this subject should be addressed alongside future prospective clinical trials. Until further data are available, no recommendations can be made on differential treatments with particle therapy in elderly patients with SGCs.

The use of systemic chemotherapy, especially cisplatin-based combinations, might face some barriers in elderly patients. The choice of systemic therapy should be independent from the chronological age, but based on performance status, comorbidities and also caregivers’ support. In the last years, as a comprehensive molecular characterization of SGCs has been developed, the therapeutic options have potentially expanded, following the principles of precision oncology. However, differently from the patients affected by high-incidence malignancies harboring analogous molecular alterations, patients with molecularly targetable SGCs are penalized by the scarcity of clinical trials specifically dedicated to these rare cancers, and they can access to innovative cancer drugs mostly through basket clinical trials, managed access programs or off-label. The type of primary tumor has to be considered in the algorithm choice; for instance, taxane and gemcitabine are not recommended in AdCC due to the lack of activity. Personalized approaches are feasible, especially for some histotypes as SDC, adenocarcinoma NOS and ca ex-PA. The search for known molecular alterations (e.g. AR and HER2 in SDC and ca ex-PA) is recommended to offer an effective therapeutic alternative to chemotherapy (e.g. ADT in case of AR overexpression; HER2-blockade in case of HER2 overexpression; NTRK inhibitors in case of NTRK fusions).

Next Generation Sequencing is the most cost-effective approach to find further actionable targets, and this technology can be available in referral cancer centers and/or by prescreening of innovative targeted therapy-oriented clinical trials. The availability of new drugs may change among countries, and the participation of fit elderly patients into clinical trials is encouraged ( Figure 3 ). Overall, the management of elderly patients with SGCs is extremely challenging and requires a complex multidisciplinary know-how. The referral to experienced facilities is recommended also for elderly patients, in order to guarantee the most accurate diagnosis and therapeutic opportunities.

EC, CVL, AZ, AJ, GG, FD, LFL, VG, VVP and LDL participated in the concept design of the manuscript. EC and LDL drafted the introduction and discussion sections, developed further by all authors. Each subsection on treatments was written by specialists of the fields: Epidemiology by GG and FD. Surgery by CVL and VVP. Radiation therapy by VG and AZ. Particle therapy by AJ. Systemic therapy by EC, LFL and LDL. All authors contributed to the revision of the final manuscript and approved the submitted version.

VVP and CVL disclosed a financial support for the publication of this article from the Walter Vandeputte Head and Neck Cancer Fund (KU Leuven, Leuven, Belgium). The remaining authors received no financial support for the publication of this article.

Author LFL received occasional conference honoraria/Advisory Board fees from Astrazeneca, Bayer, MSD, Merck-Serono, AccMed, Neutron Therapeutics, Inc. Author LDL received conference honoraria/Advisory Board fees from EISAI, MSD, Merck Serono, McCann Healthcare, Eli Lilly, Sanofi, Sunpharma, IPSEN, Bayer, Roche. Author VVP performed consultancy services for Bristol-Meyers Squibb 2019 and for ATOS medical 2020 and is a surgical trainer for thyroid surgery for Ethicon – J&J. None of these relationships interfere with the intellectual content in the current manuscript.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.